Nvidia-Backed Reflection AI Launches Beam, a 501-Billion-Parameter Open Model Matching Chinese Rivals' Reasoning Power While Cutting Inference Costs

Reflection AI, an Nvidia-backed startup, today unveiled Beam, its first frontier open-weight model, claiming reasoning performance on par with Zhipu AI's GLM-5.2 while consuming one-quarter to one-third as much inference compute. The 501-billion-parameter mixture-of-experts model will begin rolling out weights later this month under an Apache 2.0 license, marking a Western competitor to closed and open models from China's leading labs.